The overall objective of this project is to achieve a better understanding of regional brain metabolism and blood flow in epilepsy. The techniques used are the (14-C)-2-deoxyglucose and (14-C)-iodoantipyrine quantitative autoradiographic methods in animals, and the (18-F)-fluorodeoxyglucose (FDG) method with positron emission tomography (PET) in humans. The two sections of this proposal are designed to complement each other. The first portion proposes the use of FDG-PET to determine cerebral glucose metabolic rate in three categories of childhood epilepsy: Lennox-Gastaut syndrome, infantile spasms and neonatal seizures. The metabolic patterns obtained through PET will be used to assess the involvment of different brain regions in these epilepsies in order to characterize various subgroups. Subsequently, the metabolic patterns will be related to certain clinical variables, such as specific seizure type, etiology and prognosis. The information so gained will result not only in a better understanding of childhood epilepsy, but also improved management of these patients. The second section of the proposal employs animal models to address several important issues in epilepsy, that for practical reasons, are best studied in the laboratory. Two general hypotheses are presented, based on our previous work. The first hypothesis states that the phenomenon of uncoupling between hippocampal blood flow and metabolism (decreased blood flow in the presence of increased glucose consumption) during seizures may partially account for human hippocampal sclerosis. Experiments are designed to study the uncoupling phenomenon in the ictal, postical and interictal periods. Pharmacological manipulation of this phenomenon will be attempted with naloxone and physostigmine. The second hypothesis centers on the role of endogenous opioids in epilepsy. It is postulated that opioids may at least partially mediate the uncoupling phenomenon. Experiments designed to study the effects of an opioid and opioid antagonist on kindled seizures, as well as studies to achieve an animal model of hippocampal sclerosis through multiple opioid infusions which would produce repeated hippocampal hypoperfusion, are proposed.